Spaces:
Running
Running
Update app.py
Browse files
app.py
CHANGED
@@ -1,21 +1,15 @@
|
|
1 |
import streamlit as st
|
2 |
-
import torch
|
3 |
import outetts
|
4 |
from scipy.io.wavfile import write
|
5 |
|
6 |
# Initialize model configuration
|
7 |
model_config = outetts.HFModelConfig_v1(
|
8 |
model_path="OuteAI/OuteTTS-0.2-500M",
|
9 |
-
language="en"
|
10 |
)
|
11 |
|
12 |
-
#
|
13 |
-
|
14 |
-
|
15 |
-
def generate_speech(text):
|
16 |
-
with torch.no_grad():
|
17 |
-
audio, sample_rate = model.infer(text)
|
18 |
-
return audio, sample_rate
|
19 |
|
20 |
# Streamlit UI
|
21 |
st.title("OuteTTS Speech Synthesis")
|
@@ -25,9 +19,18 @@ text_input = st.text_area("Text to convert to speech:", "Hello, this is an AI-ge
|
|
25 |
|
26 |
if st.button("Generate Speech"):
|
27 |
with st.spinner("Generating audio..."):
|
28 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
29 |
output_path = "output.wav"
|
30 |
-
|
31 |
-
|
|
|
32 |
st.audio(output_path, format="audio/wav")
|
33 |
st.success("Speech generated successfully!")
|
|
|
1 |
import streamlit as st
|
|
|
2 |
import outetts
|
3 |
from scipy.io.wavfile import write
|
4 |
|
5 |
# Initialize model configuration
|
6 |
model_config = outetts.HFModelConfig_v1(
|
7 |
model_path="OuteAI/OuteTTS-0.2-500M",
|
8 |
+
language="en" # Supported languages: en, zh, ja, ko
|
9 |
)
|
10 |
|
11 |
+
# Initialize the interface
|
12 |
+
interface = outetts.InterfaceHF(model_version="0.2", cfg=model_config)
|
|
|
|
|
|
|
|
|
|
|
13 |
|
14 |
# Streamlit UI
|
15 |
st.title("OuteTTS Speech Synthesis")
|
|
|
19 |
|
20 |
if st.button("Generate Speech"):
|
21 |
with st.spinner("Generating audio..."):
|
22 |
+
# Generate speech
|
23 |
+
output = interface.generate(
|
24 |
+
text=text_input,
|
25 |
+
temperature=0.1,
|
26 |
+
repetition_penalty=1.1,
|
27 |
+
max_length=4096
|
28 |
+
)
|
29 |
+
|
30 |
+
# Save the synthesized speech to a file
|
31 |
output_path = "output.wav"
|
32 |
+
output.save(output_path)
|
33 |
+
|
34 |
+
# Play the audio in the Streamlit app
|
35 |
st.audio(output_path, format="audio/wav")
|
36 |
st.success("Speech generated successfully!")
|